Add -w to buildx ldflags

[tianon]

This disables DWARF generation, which dramatically decreases the size of the binary.

Notably, this does not include -s which disables the symbol table, as doing so would also make it so that govulncheck cannot be run on the resulting binary with meaningful results.

Before is ~90M and after is ~72M.

If we add -s as well, that ~72M drops to ~63M, but we also lose govulncheck so again, IMO, that's not worth doing.

See also:

• Preserve symbol table in binaries for static analysis (govulncheck) compose#10325
• https://github.com/docker/buildx/blob/c45185fde0c9e596afb9044be3328b6d39b3d7ca/Dockerfile#L86

[thaJeztah]

Is this something we can / should do for other binaries as well?

oh; recalls me that I tried to update to compat 13, which had a todo that an override could be removed, but that doesn't seem to work 😅

• deb: update debhelper compat level to 13 #1202

[thaJeztah]

@tonistiigi @crazy-max PTAL

[thaJeztah]

SGTM

but asked Tonis and Kevin to have a look

[tianon]

It is something you could/should do for other binaries, but I think they're all using upstream build scripts, so those changes don't happen in this repository. (For Compose in particular, this already happened back in docker/compose#10325, for example.)

[tonistiigi]

but asked Tonis and Kevin to have a look

We already do this in upstream iiuc

[tianon]

We already do this in upstream iiuc

Yep, quoted in my OP: 😅

• https://github.com/docker/buildx/blob/c45185fde0c9e596afb9044be3328b6d39b3d7ca/Dockerfile#L86

(IMO that should drop -s so that govulncheck can work on the binaries, but that's a separate discussion)

[thaJeztah]

We already do this in upstream iiuc

I'll assume that's a "LGTM", so let me merge this one.